At a minimum, shouldn't you have at least updated the unit to the current OS? Was that even an option on the demo unit? We have been working very closely with one of the companies relabeling this device. They've pretty much stripped out the BEMS app and replaced it with their own, thereby 'domesticating' the unit. Updates are pushed weekly to all units, but they need to be network-connected to receive them. At a minimum, it was suggested to network-connect the units once a month.
Sorry, I'd rather not mention the company at this time. I'm not looking to start a debate and I certainly don't want to impact our availability. Suffice to say our customers are satisfied. Overall, demand for these devices has slowed.
Have they tested with controlled elevated temperatures to see what the readings are?
To be clear, if you take this BEMS unit and you don't do any real testing, just putting people in front of it, since it pretty much always reports ~98°F and since pretty much everyone is ~98°F, people will be satisfied, albeit with a false sense of security.
Some of these units utilize facial algorithms and servers which are blacklisted by US Government. If they are deployed in a medical setting that would be an instant HIPAA violation. Those can get pretty spendy pretty quick. Hopefully your engineers did a low level check.
They make several variations with different FOV. I've asked IPVM to remove the PCB interface board and look at the PCB from front facing. Should be TO-39 can (from Heimann). Will be a HTPA32x32DR2 but whether L2.1 or L5.0 is my question.
What i mean is to present to the device your picture taken from your smartphone, wearing a mask. If you do this, the device will recognise a face wearing a mask and will sample the temperature of the smartphone. It's usually a Granted Access. This means you can enter protected premisses while on fever and without mask, as long as you carry your smartphone with you with a picture of you wearing a mask. Do that test, and see it for yourself.
Hik fails, ZK fails, Herospeed (i tested it rebranded as CDVI) fails. On the latest software version, Seek Thermal also fails (though it does not have the mask recognition feature).
To be sure you are showing the backside of the thermal sensor I think. If you remove the PCB I think you will find actual sensor facing out thru the housing and its form factor is a TO-39 can. Can you verify for me?
Posted below is the image you requested. Also, since BEMS did not recommend a specific mounting height, we mounted ours at ~5'4". We did test it with a "smaller" subject at ~4'10", but could have mounted it lower if need be, with taller subjects then needed to bow down.
I checked and application people say it is either 4.0 or 5.0 Heimann. They can't be sure. They are also interested in the board which they have never seen. If there are any identifying marks it would be helpful to get those.
Thanks for pointing that out Jiri. I manage that site and that was earlier in the temperature kiosk news cycle and I didn't excerpt properly. Fixed. If you see any more I missed please let me know as I try to be big supporter and proponent (and courteous proselytizer for). -- Craig